Open nonarkitten opened 2 years ago
This was partially implemented along with some mild compression. We broke the main table of 65536 opcodes into chunks of eight and then had an index into those chunks. This should now allow the table to get priority shuffled, but this part is not yet done.
This COULD still be implemented on the stub functions for opcodes more likely to be executed. It's value is specious now, though.
Now that the opcode table is split we need to rearrange the fragments to optimize cache locality. Consult the Cyclone 68000 project for a good template to use for which opcodes to group together.